Qwen3.8 27B vs Claude Haiku 3.5 vs Inkling Small
Too close to call on our weighted score (Inkling Small 68, Qwen3.8 27B 66, Claude Haiku 3.5 46). The right pick depends on what you value most.
Alibaba (Qwen)Qwen3.8 27B
66/100- ECI151.8
- Price$0.40 / $2.55
- Context262K
AnthropicClaude Haiku 3.5
46/100- ECI127.0
- Price$0.80 / $4.00
- Context200K
Thinking MachinesInkling Small
68/100- ECI150.5
- Price$0.58 / $1.44
- Context1.05M
Too close to call
It is close. Our weighted score puts them within 2 points (Inkling Small 68/100, Qwen3.8 27B 66/100, Claude Haiku 3.5 46/100), so choose by what matters most for your work: Qwen3.8 27B for raw capability, Inkling Small on price and Inkling Small for long inputs. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.
- CapabilityQwen3.8 27BCapabilities Index (ECI): Qwen3.8 27B 151.8 · Inkling Small 150.5 · Claude Haiku 3.5 127.0
- Lowest priceInkling SmallInkling Small $0.795 · Qwen3.8 27B $0.938 · Claude Haiku 3.5 $1.60 per 1M tokens (3:1 blend)
- Longest contextInkling SmallInkling Small 1,048,576 · Qwen3.8 27B 262,144 · Claude Haiku 3.5 200,000 tokens
- Widest inputsSame inputsQwen3.8 27B: Text, Images, Video · Claude Haiku 3.5: Text, Images, PDFs · Inkling Small: Text, Images, Audio
- Self-hostingQwen3.8 27B and Inkling SmallPublishes downloadable weights (Apache-2.0)
| Measure | Weight | Qwen3.8 27B | Claude Haiku 3.5 | Inkling Small |
|---|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 50% | 79 | 48 | 77 |
| Price | 25% | 49 | 38 | 52 |
| Inputs & features | 15% | 70 | 60 | 70 |
| Context window | 10% | 37 | 32 | 61 |
| Overall | 100% | 66/100 | 46/100 | 68/100 |
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | Qwen3.8 27BAlibaba (Qwen) | Claude Haiku 3.5Anthropic | Inkling SmallThinking Machines |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | 151.8 (best) | 127.0 | 150.5 |
| ECI rank | #54 of 165 (best) | #147 of 165 | #58 of 165 |
| GPQA DiamondGraduate-level science questions | — | 38.1% | 88.5% (best) |
| FrontierMath Tiers 1–3Research-level mathematics | — | — | 46.3% |
| OTIS Mock AIME 2024–2025Competition mathematics | — | 4.3% | 90.0% (best) |
| SimpleQA VerifiedShort factual questions | — | — | 19.1% |
| Price per million tokens | |||
| Input | $0.40 (best) | $0.80 | $0.58 |
| Output | $2.55 | $4.00 | $1.44 (best) |
| Cached input | — | — | $0.116 |
| Blended (3:1) | $0.938 | $1.60 | $0.795 (best) |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Median of 49 providers | Median of 3 providers | Official Thinking Machines API |
| Limits | |||
| Context window | 262,144 tokens | 200,000 tokens | 1,048,576 tokens (best) |
| Max output | 32,768 tokens | 8,192 tokens | 1,048,576 tokens (best) |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | Yes |
| PDFs | No | Yes | No |
| Audio | No | No | Yes |
| Video | Yes | No | No |
| Reasoning | Yes | No | Yeslow · medium · high · xhigh · max |
| Tool calling | Yes | Yes | Yes |
| Structured output | No | No | No |
| Availability | |||
| Weights | Open | Proprietary | OpenApache-2.0 |
| API model ID | — | — | thinkingmachines/Inkling-Small |
| API providers | 53 (best) | 4 | 12 |
| Released | Aug 14, 2026 | Oct 22, 2024 | Jul 30, 2026 |
| Knowledge cutoff | — | Jul 31, 2024 | — |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
- Qwen3.8 27B$9.10
- Claude Haiku 3.5$16.00
- Inkling Small$8.68
Which should you choose?
Which is better: Qwen3.8 27B, Claude Haiku 3.5 or Inkling Small?
It is close. Our weighted score puts them within 2 points (Inkling Small 68/100, Qwen3.8 27B 66/100, Claude Haiku 3.5 46/100), so choose by what matters most for your work: Qwen3.8 27B for raw capability, Inkling Small on price and Inkling Small for long inputs. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.
Which is cheaper, Qwen3.8 27B, Claude Haiku 3.5 or Inkling Small?
Inkling Small is cheaper at $0.58 input / $1.44 output per million tokens (official Thinking Machines API price). Qwen3.8 27B costs $0.40 input / $2.55 output per million tokens (median across 49 API providers); Claude Haiku 3.5 costs $0.80 input / $4.00 output per million tokens (median across 3 API providers). At a typical mix of three input tokens to one output token, that is $0.795 per million tokens for Inkling Small versus $0.938 for Qwen3.8 27B (1.2× as much) and $1.60 for Claude Haiku 3.5 (2× as much).
Which scores higher on benchmarks?
Qwen3.8 27B scores higher on the Capabilities Index (ECI): Qwen3.8 27B 151.8 (#54 of 165), Inkling Small 150.5 (#58 of 165) and Claude Haiku 3.5 127.0 (#147 of 165). The confidence ranges of the top two overlap (149.9–153.6 vs 147.7–152.4), so treat the gap as small.
Which is better for coding?
There are no published SWE-bench Verified results for Qwen3.8 27B, Claude Haiku 3.5 and Inkling Small yet, so there is no like-for-like coding score. On overall capability, Qwen3.8 27B leads, which tends to carry over to coding, but test on your own codebase. All three support tool calling for agent workflows.
Which has the bigger context window?
Inkling Small has the largest context window at 1,048,576 tokens, against 262,144 for Qwen3.8 27B and 200,000 for Claude Haiku 3.5. Maximum output per response: Qwen3.8 27B up to 32,768, Claude Haiku 3.5 up to 8,192, Inkling Small up to 1,048,576 tokens.
Which can read images, PDFs, audio or video?
Qwen3.8 27B accepts text, images and video; Claude Haiku 3.5 accepts text, images and PDFs; Inkling Small accepts text, images and audio. They handle the same number of input types.
Are any of these open source?
Qwen3.8 27B and Inkling Small publishes its weights (Apache-2.0) and can be self-hosted; Claude Haiku 3.5 is proprietary.
Which is newer?
Qwen3.8 27B is the newest, released Aug 14, 2026. Inkling Small came out Jul 30, 2026; Claude Haiku 3.5 came out Oct 22, 2024. Knowledge cutoff: Claude Haiku 3.5 Jul 31, 2024.
How do you decide the winner?
Each model gets a 0–100 score on capability (50%, independent benchmark results); price (25%, blended price per million tokens (3 input : 1 output), log scale); inputs & features (15%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (10%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Nov 2, 2026.
Related comparisons
- Qwen3.8 27B vs Mistral Large 4Mistral AI · ECI 150.6
- Qwen3.8 27B vs DeepSeek V4 Pro 0813DeepSeek · ECI 155.9
- Qwen3.8 27B vs Gemini 3 Flash PreviewGoogle · ECI 152.0
- Claude Haiku 3.5 vs Mistral Large 2.1Mistral AI · ECI 128.6
- Claude Haiku 3.5 vs Llama-3.3-70B-InstructMeta · ECI 127.3
- Claude Haiku 3.5 vs GPT-4oOpenAI · ECI 129.0
